Abstract: Larger Object Oriented(OO) software systems can benefit greatly from the use of software design patterns. Because design pattern offers a well-structured and proven solution to recurrent issues that developers deal with. Previous studies explored the effects of design patterns and created several models for mining design patterns. Some of them were developed using machine learning (ML) methods. This paper concentrated on giving a brief summary of the difficulties with design patterns from the developer's point of view, as well as their impact on some quality aspects. Along with that, we discussed the shortcomings of contemporary classic pattern mining techniques. We compared both conventional mining tools and ML-based mining tools. Findings show that both the random forest model and PatternScout, which is based on SPARQL queries, can effectively discover all patterns with higher precision and recall values.
External IDs:dblp:conf/segah/RahmanCS23
Loading